[Bugfix][Scheduler] Let pooling chunked prefill use full context (#48039)
Signed-off-by: Tommy <tran.tommy@hotmail.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> Co-authored-by: OpenAI Codex <codex@openai.com> Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
co-authored by
OpenAI Codex
Harry Mellor
parent
cbee82324c
commit
e6fc81bc78
@@ -18,6 +18,7 @@ from vllm.config import (
|
||||
SpeculativeConfig,
|
||||
VllmConfig,
|
||||
)
|
||||
from vllm.config.model import RunnerOption
|
||||
from vllm.config.scheduler import SchedulerPolicy
|
||||
from vllm.multimodal.inputs import (
|
||||
MultiModalFeatureSpec,
|
||||
@@ -78,6 +79,7 @@ def create_scheduler(
|
||||
ec_role: str | None = None,
|
||||
use_v2_model_runner: bool | None = None,
|
||||
kv_cache_spec: KVCacheSpec | None = None,
|
||||
runner: RunnerOption = "auto",
|
||||
per_request_spec_decode_metrics: str = "none",
|
||||
scheduling_policy: SchedulerPolicy = "fcfs",
|
||||
diffusion_canvas_length: int | None = None,
|
||||
@@ -109,6 +111,7 @@ def create_scheduler(
|
||||
# The scheduler reads model_config.max_model_len, not the
|
||||
# SchedulerConfig one, so both must agree.
|
||||
max_model_len=max_model_len,
|
||||
runner=runner,
|
||||
)
|
||||
if mm_encoder_only:
|
||||
model_config.multimodal_config.mm_encoder_only = True
|
||||
|
||||
Reference in New Issue
Block a user